@@ -19,7 +19,7 @@ on testing scope.
19
19
or taken prebuilt from [ releases] ( https://github.com/intel/llvm/releases ) .
20
20
- LIT tools (llvm-lit, llvm-size). They are not available at prebuilts above,
21
21
but can be built in compiler project (e.g. with "ninja check").
22
- - Target runtime(s) to execute tests on devices other than host . See
22
+ - Target runtime(s) to execute tests on devices. See
23
23
[ installation instructions] ( https://github.com/intel/llvm/blob/sycl/sycl/doc/GetStartedGuide.md#install-low-level-runtime )
24
24
25
25
# Build and run tests
@@ -40,7 +40,7 @@ With compiler tools available in the PATH:
40
40
cmake \
41
41
-DCMAKE_CXX_COMPILER=clang++ \
42
42
-DTEST_SUITE_SUBDIRS=SYCL \
43
- -DCHECK_SYCL_ALL="opencl:host " \
43
+ -DCHECK_SYCL_ALL="opencl:cpu " \
44
44
..
45
45
46
46
# Build and Run
@@ -56,7 +56,7 @@ cmake \
56
56
-DCMAKE_CXX_COMPILER=clang++ \
57
57
-DTEST_SUITE_SUBDIRS=SYCL \
58
58
-DSYCL_BE="opencl" \
59
- -DSYCL_TARGET_DEVICES="host " \
59
+ -DSYCL_TARGET_DEVICES="cpu " \
60
60
..
61
61
62
62
# Build and Run
@@ -96,7 +96,7 @@ list of configurations. Each configuration includes backend separated
96
96
from comma-separated list of target devices with colon. Example:
97
97
98
98
```
99
- -DCHECK_SYCL_ALL="opencl:cpu,host ;ext_oneapi_level_zero:gpu,host ;ext_oneapi_cuda:gpu;ext_oneapi_hip:gpu;ext_intel_esimd_emulator:gpu"
99
+ -DCHECK_SYCL_ALL="opencl:cpu;ext_oneapi_level_zero:gpu;ext_oneapi_cuda:gpu;ext_oneapi_hip:gpu;ext_intel_esimd_emulator:gpu"
100
100
```
101
101
102
102
*** SYCL_BE*** - SYCL backend to be used for testing. Supported values are:
@@ -108,11 +108,10 @@ from comma-separated list of target devices with colon. Example:
108
108
109
109
110
110
*** SYCL_TARGET_DEVICES*** - comma separated list of target devices for testing.
111
- Default value is cpu,gpu,acc,host . Supported values are:
111
+ Default value is cpu,gpu,acc. Supported values are:
112
112
- ** cpu** - CPU device available in OpenCL backend only;
113
113
- ** gpu** - GPU device available in OpenCL, Level Zero, CUDA, and HIP backends;
114
114
- ** acc** - FPGA emulator device available in OpenCL backend only;
115
- - ** host** - SYCL Host device available with all backends.
116
115
117
116
*** OpenCL_LIBRARY*** - path to OpenCL ICD loader library. OpenCL
118
117
interoperability tests require OpenCL ICD loader to be linked with. For such
@@ -157,7 +156,7 @@ If UNSUPPORTED:sycl-ls specified, test will run only if sycl-ls tool is
157
156
unavailable.
158
157
159
158
* ** windows** , ** linux** - host OS;
160
- * ** cpu** , ** gpu** , ** host ** , ** accelerator** - target device;
159
+ * ** cpu** , ** gpu** , ** accelerator** - target device;
161
160
* ** cuda** , ** hip** , ** opencl** , ** level_zero** , ** esimd_emulator** - target
162
161
backend;
163
162
* ** sycl-ls** - sycl-ls tool availability;
@@ -182,8 +181,7 @@ Following options can be passed to llvm-lit tool through --param option to
182
181
configure specific single test execution in the command line:
183
182
184
183
* ** dpcpp_compiler** - full path to dpcpp compiler;
185
- * ** target_device** - comma-separated list of target devices (cpu, gpu, acc,
186
- host);
184
+ * ** target_device** - comma-separated list of target devices (cpu, gpu, acc);
187
185
* ** sycl_be** - SYCL backend to be used (opencl, ext_oneapi_level_zero,
188
186
ext_oneapi_cuda, ext_oneapi_hip, ext_oneapi_intel_emulator);
189
187
* ** dump_ir** - if IR dumping is supported for compiler (True, False);
@@ -220,7 +218,7 @@ configure specific single test execution in the command line:
220
218
Example:
221
219
222
220
```
223
- llvm-lit --param target_devices=host, gpu --param sycl_be=ext_oneapi_level_zero \
221
+ llvm-lit --param target_devices=gpu --param sycl_be=ext_oneapi_level_zero \
224
222
--param dpcpp_compiler=path/to/clang++ --param dump_ir=True \
225
223
SYCL/External/RSBench
226
224
```
0 commit comments